Simultaneous Computation of Two Independent Tasks Using Reservoir Computing Based on a Single Photonic Nonlinear Node...

Area of Science:

• Photonics
• Optical Computing
• Reservoir Computing

Background:

• Reservoir computing offers a novel approach to complex computations.
• Photonic systems provide high-speed processing capabilities.
• Simultaneous processing of multiple tasks is a key challenge in computing.

Purpose of the Study:

• To demonstrate a photonic delay-based reservoir computing system capable of parallel processing of two independent tasks.
• To utilize a single semiconductor ring laser (SRL) with optical feedback for parallel computation.
• To mitigate crosstalk between tasks by employing the two directional optical modes of the SRL.

Main Methods:

• Numerical demonstration of a photonic reservoir computing system.
• Utilizing a single-longitudinal mode semiconductor ring laser (SRL) with optical feedback.
• Analyzing performance on chaotic time series prediction and nonlinear channel equalization benchmark tasks.

Main Results:

• The system successfully processes two independent computational tasks in parallel using two directional optical modes of the SRL.
• Good performance was observed for simultaneous prediction/classification tasks.
• Slight performance degradation was noted due to nonlinear and linear interactions between the modes, but overall good performance across a broad parameter range was achieved.

Conclusions:

• The proposed photonic reservoir computing system is feasible for parallel processing of independent tasks.
• The use of directional modes in the SRL effectively mitigates crosstalk.
• The system demonstrates potential for efficient and simultaneous handling of diverse computational challenges.